Review Pack
Decision narrative, risks, forces, ADRs, and implementation plan.
Turn one MVP brief into an agreed architecture, dangerous assumptions, and a build order your coding agent—or engineer—can actually follow.
12
components
18
typed flows
03
open risks
One semantic model
Every artifact points to the same system.
Trade-offs exposed
Chosen and rejected options stay visible.
Risks before files
Auth, data, payments, and failure modes.
Structurally checked
Coverage and consistency—not vague confidence.
/ 01 · Inspect the system
The canvas is one view. Risks, API boundaries, data relationships, deployment, and build order all resolve back to the same typed model.
One source · five views · explicit trade-offs
A Spire package keeps diagrams, API contracts, ERDs, flows, deployment, critique, and validation tied to the same underlying model.
/ 02 · The package
Four connected evidence layers. Each one answers a different question before implementation begins.
Decision narrative, risks, forces, ADRs, and implementation plan.
Typed components, edges, storage, queues, gateways, and service boundaries.
ERD, API contract, sequence flows, and deployment topology.
Coverage, orphan checks, critique findings, and confidence metadata.
/ 03 · Review sequence
Spire is intentionally not a blank canvas. The sequence keeps product intent, architecture choices, critique, and structural checks connected.
Users, critical journeys, constraints, integrations, and what can wait.
Boundaries, data ownership, APIs, deployment, and meaningful alternatives.
Assumptions, dangerous blind spots, rejected options, and unresolved risks.
Journey coverage, identifier integrity, reachability, and cross-artifact consistency.
What leaves the review
Trace
See how the package was assembled.
Critique
Challenge risks and assumptions.
Export
Move the agreed plan into the repo.
Why Spire
Diagram tools help you draw and general AI helps you discuss. Spire starts from the product brief, builds one semantic model, and keeps its architecture views and decision narrative connected.
| Workflow | Spire | Diagram workspace | General AI chat |
|---|---|---|---|
| Starting point | A guided product brief | A blank canvas or existing text | One prompt at a time |
| Primary output | A connected architecture package | An editable visual | A conversational answer |
| System, data, APIs, and deployment | Derived from one semantic model | Kept aligned by the author | Generated across separate responses |
| Risks, assumptions, and rejected options | Included in the review pack | Added manually | Depends on every prompt |
| Structural checks | Coverage, reachability, and consistency | Visual or syntax validation | Verified manually |
| Buying model | $19 once for one architecture | Usually recurring tool access | Separate AI subscription or usage |
Use the public sample to inspect the workflow before checkout. Personalized architecture generation begins only after a Single Architecture payment or another paid access offer.
/ 05 · Start with one build
Begin with one Single Architecture. Repeat-access options exist for builders who already know they need another project or their own model routing.
$19
once
For a founder or engineer preparing one real product build.
One payment · no subscription
Create one for $19 →$9
/ 30 days
For a solo builder planning one product with their own model key.
Provider bills usage and controls balance, quotas, rate limits, and availability
Choose Builder BYOK →$200
once
For repeat builders who already have a compatible AI provider key.
AI provider usage is billed separately
Choose Founder BYOK →$29
/ 30 days
For an active build that needs managed generation without key setup.
A fixed 30-day pass · does not auto-renew
Choose a Build Month →Purchase one architecture, run the generation, then open the complete diagrams, documents, and exports in your workspace.